November 1970 Popular Electronics
[Table of Contents]
People old and young enjoy waxing nostalgic about and learning some of the history of early electronics. Popular
Electronics was published from October 1954 through April 1985. All copyrights are hereby acknowledged. See all articles from
of the least known but most fascinating of semiconductor devices
is the light-emitting diode (LED). Until quite recently, these devices
were too expensive for widespread use." That statement is hard to
imagine in 2011. It was written in late 1970, just 40 years ago.
Now, LEDs seem to be in every consumer device, whether it be a simple
power ON indicator light or an array of alpha-numeric displays.
TV and stereo remotes use infrared LEDs, optoisolators that use
LEDs are integral components of modems, motor controllers, and motion
sensors. Commercial truck tail lights are made of LED clusters,
as are the grotesquely expensive LED bulbs that are verrrry slowly
replacing incandescent and the toxic (mercury) CFL household bulbs.
What was once rare is now a commodity - it happens all the time.
See all articles from
Light Emitting Diodes
New Semiconductors for Readout and Communication
Cover story by Forrest M. Mims, III
of the least known but most fascinating of semiconductor devices
is the light-emitting diode (LED). Until quite recently, these devices
were too expensive for widespread use; however, technological advances
in their fabrication now make possible moderately low prices so
that they are attractive to the electronics experimenter.
The first recorded instance of light being generated by a "diode"
was in 1907 when H. J. Round touched a pair of battery wires to
a crystal of silicon carbide. Much to his surprise, flashes of yellow
light were emitted at the contact region of one of the battery wires
- he had accidentally discovered the LED. Unfortunately, his discovery
was forgotten and not until the early 1950's did scientists once
again study semiconductor light emission. At that time several patents
were applied for covering LED's made from silicon or germanium -
common semiconductor materials. One of these patents not only described
the principle of the LED but listed several fascinating uses for
the device. Among them were light-beam communication systems, light
"radar", and light beam alignment devices.
Fig. 1. Most infrared LED's emit light at a wavelength of about
0.9 microns of the electromagnetic spectrum. Visible LED's may
emit from 0.53 microns to the limit of visible region at about
Unlike the 1907 silicon carbide cat whisker diode, the 1950's LED's
emitted infrared light (see Fig. 1). The invisible beam was desirable
but researchers began a concentrated effort to fabricate LED's in
the visible range. The first company to market any kind of LED was
Texas Instruments, Inc. Their diodes included infrared emitters
made from gallium arsenide (GaAs) and visible emitters made of gallium
arsenide phosphide (GaAsP). These diodes were expensive; but their
appearance on the commercial market in 1962 whetted the appetites
of design engineers - and of course was of great interest to other
semiconductor manufacturers. While scientists at IBM and Bell Telephone
performed basic research on the devices (especially visible emitters),
General Electric, Monsanto, Electro-Nuclear Laboratories and others
began competing with TI.
Since the LED has an almost unlimited
lifetime and because of its low operating current, many claims were
made for its potential in flat screen television, indicator lamps,
night lights, and even as a source of room lighting. The extremely
fast modulation capability of the LED made possible several demonstrations
of voice communications permitting two parties to converse over
a beam of invisible infrared light for clear weather distances
of several miles. How Does It Do It?
The LED is different from conventional incandescent lamps because
the latter give off light as a byproduct of heat. That is, the filament
must first be heated to incandescence before light is produced.
This, of course, is the reason for placing incandescent filaments
in evacuated or gas-filled glass bulbs. If the filament were exposed
to oxygen, it would quickly be consumed. Not so with LED's; they
operate with or without the presence of air. In fact, the main reason
LED's are placed in containers is to protect the rather delicate
contact wires. Often, in fact, the LED is simply protected by a
layer of clear epoxy which serves both as a protective cover and
a lens. Since the LED produces far more light than heat, it is a
more efficient source of light than the incandescent lamp.
Production of light by an LED comes under the broad heading
of "electro-luminescence." Luminescence describes light produced
by means other than incandescence. Since luminescing bodies are
generally at the temperature of their environment, they are sometimes
referred to as sources of "cold light." Fireflies, many species
of fish, rotting wood, marsh gas, and many other objects are sources
of luminescent light.
Light from an LED is the result of
stimulation by a small electric current. As is the case with all
luminescent bodies, the light is a result of sub-atomic events.
The cause of these events may be chemical, as with the firefly,
or electrical as with the LED. Light Generation
In the LED, emission of light is called "PN junction luminescence."
As shown in Fig. 2, light is emitted at the junction of an LED when
electrons which have been stimulated to higher than normal energy
levels move across the junction and fall back into their normal
places. The energy loss resulting when an electron reoccupies its
normal slot in an atom is accompanied by the emission of a photon
of light at a wavelength related to the difference in energy between
the two bands. Since junction luminescence involves the combination
of electrons with holes, phvsicists often refer to the effect as
Fig. 2. Light is emitted when an electron crosses the np junction
interface and falls from the excited to the unexcited state.
Fig. 3. The critical interface problem. Much more light is emitted
by hemispheric diode but some is lost in the thick n-region.
Some planar emitters are provided with a reflector to give greater
The sub-atomic physics of semiconductor light emission can be highly
efficient. But the mechanisms of bringing the light to the surface
and the application of electrical current to the diode can be terribly
inefficient. Sources of inefficiency are:
1. Internal absorption
of light within the p or n region of the semiconductor following
its emission at the junction.
2. Reabsorption of light which
is reflected back into the diode due to the "critical interface"
3. Resistance at the electrical contacts and in
Many of these inefficiencies have been
partially resolved as a result of years of research. Consider, for
example, the critical interface problem. Since most semiconductors
have a high index of refraction, they tend to reflect light at air-semiconductor
interfaces. The effect is similar to that of an observer looking
through a glass window and seeing his own reflection as well as
objects on the other side of the window. The effect results in the
loss of light that is generated at the junction.
solution to the problem was suggested by early research and was
first used commercially by Texas Instruments. It consists of forming
the light emitting region of the diode into a dome. Since the light
is only bent and not reflected back into the diode, all light reaching
the surface is emitted (see Fig. 3). Unfortunately, the dome approach
does have its drawbacks - the longer path that the light must travel
before being emitted causes more internal absorption than in planar
emitters. Also, grinding domed diodes is an expensive process.
In an effort to gain the benefits of both planar and domed
diodes, many manufacturers now coat planar diodes with a transparent
dome of epoxy which essentially serves the same purpose of the semiconductor
dome. PHYSICS OF LIGHT EMITTING DIODES
Modern atomic physics tells us that electrons within the
confines of a particular atom are allowed to occupy only discrete
energy levels or bands. Energy levels between the outer two bands,
the valence and conduction bands, are separated by a region called
the forbidden gap. Under special conditions of doping, an electron
may occupy one or more levels within the forbidden gap for relatively
brief periods of time. The forbidden gap plays an important role
in semiconductor light emission, since transitions from the conduction
to the valence hand provide the mechanism for photon generation.
efficient way to cause transitions between the valence and conduction
bands is to pump or inject electrons into the n region of a semiconductor
diode. If a sufficient number of electrons is injected, the potential
barrier at the junction of the n and p regions will be overcome
and current will flow through the diode. Having crossed the junction
barrier, the injected electrons seek their equilibrium point and
drop from their excited position in the conduction band to the valence
band. In essence, the electrons drop into holes or regions in a
band where there is an electron deficiency.
The act of electrons
combining with holes is called recombination and results in the
magic of PN junction electroluminescence or light emission. An electron
falling from a high to a low level releases the energy which propelled
it over the junction as heat, light, or some of each. Heat emission
is accompanied by vibrations in the crystal lattice and if left
uncontrolled, results in thermal destruction of the diode. Heat
emission predominates in indirect band gap semiconductors such as
silicon and germanium. The forbidden gap in such materials permits
an electron falling from the conduction to the valence band to loiter
briefly at one or more levels. Transitions between the various levels
in the forbidden gap may result in either light or heat emission.
It is interesting to note that an ordinary silicon or germanium
diode emits a minute quantity of infrared light when forward biased.
However, light emission in such diodes is far less efficient than
heat production and development of practical semiconductor light
emitters followed research with direct band-gap materials.
Gallium arsenide GaAs is normally a direct band-gap material.
Since electrons injected into a GaAs diode recombine with holes
without pausing at intermediate levels within the forbidden gap,
light emission is very efficient and production of heat is generally
limited to contact resistance and bulk absorption of some of the
emitted light within the semiconductor.
GaAs diodes are
so efficient as light emitters that scientists chose them as candidates
for early work concerning the feasibility of fabricating semiconductor
lasers. In the fall of 1962, researchers at G.E., IBM, and MIT announced
almost simultaneously lasers made from specially prepared GaAs diodes.
In a future issue of POPULAR ELECTRONICS, the operating principles
and characteristics of the diode laser will be described.
How Are They Made?
As mentioned earlier,
the LED usually consists of a PN junction of one type of semiconductor,
the most common being gallium arsenide. A typical diode may consist
of a small wafer of positively charged (p) material in intimate
electrical contact with a metal header similar to a transistor case.
Before it is mounted on the header, the wafer is given a thin top
layer of negatively charged (n) GaAs by either diffusion or epitaxial
growth. The P portion of the finished wafer is connected to one
of the wire leads of the header, resulting in a true PN junction
To collect the light output from the wafer in an
efficient manner, a metal can with a lens or flat window at one
end is welded to the header. The can also serves as protection (see
How Are They Used?
Fig. 4. In practical applications, the LED is mounted within
a metal enclosure, similar to a transistor, for mechanical protection.
A transparent window permits the light to exit from one end
of the enclosure.
Now that we know something about
the history, physics, and mechanical construction of the LED, what
are its applications other than those found in the laboratory? Especially,
how can the experimenter use these tiny sources of "cold light."
An important military use of the LED is as a covert source
of illumination for night vision devices. The usefulness of the
Army's starlight scope is greatly enhanced by invisible infrared
illumination supplied by one or more LED's. Another military use
is in covert communications. The Navy, for example, has several
types of infrared, line-of-sight LED voice communicators for ship-to-ship
Recently, commercial applications of
the LED have been revealed at an unprecedented rate. Several companies
are mass producing tiny alpha-numeric displays consisting of arrays
of LED's. These displays operate at a much lower voltage than conventional
mechanical, incandescent, cathode ray, or gas-discharge displays.
One manufacturer has demonstrated a complete VOM in a probe!
Using several of the new numeric displays, the futuristic unit is
similar in size to a standard scope probe. Volts and ohms are read
directly from the small light display. The Hamilton Watch Company
will market a digital watch using LED numeric displays, and Bell
Telephone is hard at work on new types of blue, green, yellow, and
red LED's for use in home telephones. Infrared LED's are being used
in several types of new intrusion alarms. One company even markets
a complete line of LED voice communicators and burglar alarms.
One of the most unusual applications of the infrared LED
is as a light source for small mobility aids for the blind. Such
devices may eventually be marketed at a price comparable to most
An important result of all these new uses
for LED's is a large drop in price. LED's are available for under
$2.50 each in small quantities. This price is competitive with that
of miniature, long-life indicator lamps. And of course, the LED
offers sturdier packaging, a million-hour lifetime, and far less
Only a year ago, the cheapest infrared
LED retailed for $18.00. At least five manufacturers now offer GaAs
LED's for under $7.50. These devices are far more efficient than
visible LED's and are therefore usable in experiments in secret
To sum up then, the LED is superior to the
conventional filament lamp in the following ways:
time is extremely fast most LED's operate in a fraction of a second
rise and fall time . Some LED' can reach a speed of 100 MHz.
2. Because it is a solid-state device, an LED has no warm-up
time, is completely free of microphonics, is just about impervious
to mechanical vibration and other environmental conditions, and
is usually of very small size and weight.
3. The LED light
output is nearly monochromatic-far from being a laser but close
enough to have most of its light in a relatively narrow bandwidth.
This makes it possible to use optical filters to reduce ambient
4. The LED is a low-impedance device with forward
characteristics similar to those of a conventional silicon diode
and it can be driven from ordinary low-voltage supplies with conventional
Elsewhere in this issue you will find
a unique construction article for a low-cost LED communicator. Other
projects using LED's will appear in future issues.
For more information about
commercial light emitting diodes, write to one of the following
- Electro·Nuclear Laboratories 115 Independence Drive Menlo
Park, CA 94025
- Miniature Lamp Department General Electric Company Nela
Park Cleveland, OH 44112
- Monsanto Electronic Special Products 10131 Bubb Road Cupertino,
- Radio Corporation of America Electronic Components and Devices
Harrison, NJ 07029
- Texas Instruments Semiconductor Components Division Box
5012 Dallas, TX 75222
- Fairchild Semiconductor 313 Fairchild Drive Mountain View,
- Hewlett-Packard 1501 Page Mill Road Palo Alto, CA 94304
- Motorola Semiconductor Products Box 20912 Phoenix, AZ 85036
- Sharp Electronics Corporation 178 Commerce Road Carlstadt,
- An excellent booklet on LED's is sold by the General Electric
Company for $2.00. Called the "Solid State Lamp Manual," it
describes in detail theory, characteristics, and applications.